Halo Wars 2 Benchmarks [Computerbase]

[Bacon1]

https://www.computerbase.de/2017-02/halo-wars-2-grafikkarten-benchmark/

 

[Mondozei]

It's noteworthy that the TXP is the only GPU which can breach 60 fps in 4K. Furthermore, If you compare the jump from the 1080 to the TXP, it's just 1% in 1080p. it then jumps a massive 22% in 1440p and then finally goes up to 28% in 4K.

I wonder if it's just the VRAM which is the difference or if it's something more.

P.S. Did they stop benchmarking the 980 Ti? CB usually had a very large selection of GPUs in their testing, second only to TPU. I guess one can always take the GTX 1070's results and just multiply it by 0,95 to get an guesstimate, but still, somewhat surprising if they just stopped. It's a widely used card to this day.

 

[Mondozei]

Another weird thing. Compare the Fury X with the 1070. It's 13% difference in favor of the 1070 at 1080p. And it increases that advantage to 19% at 1440p.

But by the time you get to 4K, the FuryX not only completely closes the gap, it even goes ahead of the 1070 by 2%(!). Very strange. I know that the Fury has very good color compression, but the 1070 is part of a newer uarch and it has more VRAM.

Also, if Fury scaled better with higher resolution, you would have seen the gap close by 1440p compared to 1080p, instead it widened to then totally close and even end up with a slight advantage for Fury at 4K. I can't make any sense of it. Anyone has any ideas?

 

[Mercennarius]

AMD Cards tend to be doing better at increased resolutions relative to NVidia cards. 390X and RX 480 go from being slower than the 980 and 1060 at 1080P to ahead of the 980 and matching the 1060 at 4K.

 

[SlickR12345]

The Fury X is bottlenecked by the limited pixels and feeding those shaders. At 4k resolution pure RAW performance from those 4000+ shaders is able to win out against the more efficient GTX 1070.

Good to see the 1060 winning a DX12 title, it is usually same performance or lower with AMD counterparts, but good to see that it can absolutely win in this field.

 

[Krteq]

The Fury X is bottlenecked by the limited pixels and feeding those shaders. At 4k resolution pure RAW performance from those 4000+ shaders is able to win out against the more efficient GTX 1070.

Can you please elaborate?

What do you mean by "limited pixels"? A pixel fillrate? It doesn't make sense.

Heh, "absolutely win" ... seriously?

 

[tamz_msc]

It's an RTS so it will be CPU bound at anything less that 4K. Add to that the CPU driver overhead of the Fury X at lower resolutions, which is why the Fury X can close the gap only when you crank up the pixel count.

 

[SlickR12345]

The game is terrible though, bad and loose controls, units with a mind of their own, sloppy movement, weird pathfinding and just generally crappy console gameplay. This is why RTS doesn't work on consoles, coz its bad. Controllers are not intuitive and then porting controllers focuses gameplay to mouse and keyboard makes the game feel bad.

Bad as far as graphics and performance, its a very good game.

 

[nathanddrews]

TitanXP is 3X faster than my 970 at 4K. 1070/Fury are 2X faster... ouch.

Curious to know how Ryzen compares to the 6700K in this test. I have to believe that this game is well threaded, but I'd like to see. I think I see a Ryzen+Vega/Volta upgrade in my future...